Minutes of EXOGAM Design Meeting
Agata/AMBMin/Oct07/final
Minutes AGATA Management Board
Next Meeting: 25th October 2007 start 9:00 finish 17:00. University of Cologne.
Present:
D. Bazzacco
A. Gadea
G. Duchêne
J. Nyberg
P. Reiter
J. Simpson Project Manager and Secretary
Ch. Theisen
W.KortenASC representative
1. Apologies for absence
None
2. Minutes of the last meeting16-17th July GSI.
The draft minutes Agata/AMBMin/Jul07/02D were accepted as a correct record of the meeting.
Will distributed as Agata/AMBMin/Jul07/final.
3. Matters arising from the last meeting and actions outstanding from last phone conference
The final technical report on AGATA was discussed. This document is stated as a requirement of the MoU and as part of the EURONS project.
Action JS to initiate preparation by the AMB of the AGATA technical report (end of 2008). Need to check on MoU requirements (current and future).
Action PR: Report on the first triple module is required (for EURONS) by end of this year.
Action: Web page discussion JN could be responsible with help from students. Saclay are offering a person to contribute to the web design? Both options seem good so will discuss at next phone conference or when WK, JN and JG are together.
Ongoing
Actions on the AMB from the ASC meeting:
- AMB to formalise responsibilities for the above and to provide a planning for the commissioning experiments.
Action on JN and AG to prepare a plan, see below.
- AMB to contact A. Korichi to sort out how this is to be done, in agreement with the EU tendering regulations.
Done.
- AMB to investigate needs and costs for data storage beyond the budgeted 100 Tb.
- An evaluation of the required human resources will be made by the AMB, W.K. and J.S
- Running cost analysis
Action all: To consider what effort will be available, resources and tasks that each country and laboratory will contribute to the AGATA 1.
Initial feedback to be reported at the next AMB.
Ongoing.
Action all, in addition to comments above any extra information or “real costs” to be sent to WK.
Ongoing
Effort requirements still required.
Action All: All information by 23rd November (Capital and effort for the new MoU to 2012).
HV controller units: New layout being designed at Saclay with connector mounted.
Action all. Check what has been ordered from CTT (re HV).
Italy has checked and its contract says:
…The cryostat will be equipped with three high voltage supplies (ISEG MHx 50 103 2 2.5), supply via D Sub 15 housing mounted plugs, or with high voltage filters and standard SHV plugs. (remark 3) …
Remark 3: Still to be settled is the HV supply of the AGATA experiment. CTT suggested and thereby quotes to equip the cryostats with individual ISEG HV modules mounted directly to the cryostat housing. As an alternative CTT accepts to equip the cryostat with HV filters and standard SHV plugs for the same price.
Italy: ISEG is explicitly written but is it ISEG/SHV plugs?.....
Germany similar
UK: HV supplies included.
Sweden/Turkey Plugs or modules provided by customer.
Possibility of recording on tapes?
Action: CT to investigate if tape backup is feasible. Cost of drives etc.
Done see DAQ report.
Grid? Need a virtual AGATA organisation for test.
Action: JN to investigate in detail what is possible and what’s involved.
See DA report.
Action DB to produce a list of required electronics.
Remove until decision for an in-beam test is taken.
Action: JN to co-ordinate the simulation and which reactions could be used.
Remove until decision for an in-beam test is taken.
Action: All to consider the detailed commissioning experiments at the next AMB.
Paper to be prepared by AG and JN. To be circulated to the AMB prior to the AGATA week.
The plan will be presented at the next AGATA week.
Ongoing.
4. ASC report
Next meeting 4th December in Cologne.
FP7 status. AGATA bid was not successful. Funds will be integrated into host laboratories access budget. Hopefully this will be additional funding. Allocation to AGATA has to be agreed with host laboratories.
An MoU/agreement between AGATA and each host laboratory will be agreed. Timescales for this are to have the MoU with Legnaro in place by early 2008 and an MoU for Ganil in place by early 2009. These MoUs will specify what the host laboratory and the wider AGATA collaboration provides. Operation costs must be provided by the collaboration starting at the latest in 2009.
The new MoU is being prepared. It is expected that this new MoU will commence on 1/1/2008. It is possible that the majority of the stakeholders will have signed this MoU by spring 2008 but this will be discussed again at the ASC. This date takes into account the financial and decision making process in each country.
5. Funding status and capital allocation
Action: All to check GD proposed redistribution of finances.
Check latest version of capital file, GD and all to look for latest and circulate.
Next iteration could possibly swap funds in Sweden for DAQ to DA?
TUM has ~15k euro. What could this money be used for?
GD to discuss with S.Gales how Amel/ IPNO could proceed to obtain the funds for pre-processing.
6. Reports from Working Groups
- Detector module and characterisation
Report from PR is attached.
The AMB agreed that if the fixes of the microphonics do not work then the detectors should be rejected.
Saclay will pay for the replacement dewar on their test cryostat (vacuum leak).
C001 in Munich cryostat or C003 in Liverpool cryostat will be sent to Liverpool for scanning ASAP after the faults given in the report are fixed.
Questions again regarding the cost of triple cryostats!
Synchronisation of GRT4 clocks is an issue? Discuss at AGATA week.
GSI scanning system. Overestimation of data rate. Have some data which is to be compared with Liverpool data. More details needed which will be reported at the AGATA week.
Orsay scanning system. The present setup works with the GABRIELA digital electronics which is used for ~2 months periods in Dubna. Upgrade foreseen for a fixed digital electronics.
Orsay is ready to get a detector by Feb 2008.
Detector group needs a new team leader.
Candidate Bart Bruyneel (Cologne post doc).
This was agreed.
Action PR to inform Bart.
b. Front end processing
From DB’s presentation.
Digitization
• Prototype delivered to LNL at the beginning of August
– to define final parameters (gain, bandwidth)
– tested but no real work done
• Construction of 18 units proceeding well
– Modules being tested
– Boards will need to be modified
• for final gain parameters and Nyquist filter
• for correct handling of ToT
• for new core preamplifier with two ranges
– One order still in discussion (Turkey)
Possibility of buying the 3 digitizers in the UK.
Status of Turkey to be ascertained. Turkey could the funds be used for something else
Action JS
Pre-processing
• New ATCA Carrier
– 3 PCBs available 2 prototype mounted
– hardware tests ~80 % ready
• DPRAM soldered with an offset in one of them
– revised version
• November/December: new routing (CERN) and PCB production
• January/February: population and tests
• New segments Mezzanine
– 3 prototypes mounted
– hardware tests ~completed
– no revision foreseen
• New Core Mezzanine
– ready for PCB production
– awaiting results of integration tests
• New GTS Mezzanine
– on hold
• Integration tests of Carrier+3 Segment mezzanines + 1 GTS (1st version)
– ongoing at Padova (22-31 October)
• Ch.Oziol (IPNO) + B.Travers and X.Lafay (CSNSM)+ Padova Team + I.Lazarus (few days)
– testing optical fibers input (MGTs in loopback)
– readout to carrier DPRAM
The AMB wishes to congratulate the effort so far in the production of the new hardware.
• Production HW (assuming integration tests successful)
– Carrier
– ready for production according to plan (mid March 2008)
– production (36+ boards) to be managed by IPNO
• Enough know how? (idea is to have the 3rd PCB tested by Ch.Oziol+ in Padova)
• enough manpower?
• Boundary scan to ease testing (RAL)
• Test and validation ~1 week/board/person
• Availability of funds?
DB will discuss IPNO effort etc. with J.Pouthas.
New engineer (3rd) will start in Orsay.
– Segment Mezzanine
– ready for production according to plan (end March 2008)
– production (~108+ boards)
• Test and validation x-days/board/person
• Availability of funds?
– Core Mezzanine: present plan is beginning of July 2008?
– GTS: present plan is begin February 2008 not realistic
Need to decide what are the long delivery time items that are safe to order and attempt to place orders for long delivery items as soon as possible.
• Production VHDL for Carrier
– is the critical issue typically ~70 % of the effort
– integration of code estimated to take 8 month to a skilled person
– discussions on how to proceed:
• IPNO alone (as planned) or Padova still involved
• choice of RTOS for the FPGA (VxWorks or Linux) depend on who does the work
• main issue is the availability of man power to develop the VHDL
Need a proposed solution and confirmation or not from I Lazarus. Ideally need a clear path by the AGATA week, but at the latest for the next ASC meeting
• Implementation of processing algorithms (Segment and Core Mezzanine.)
– IHPC code implemented in segment mezzanine as a black-box
– Plans to develop the Strasbourg code using a GTS mezzanine
• Base Line restorer
• Time over threshold
• Readout of carrier
– PCI Express (RAL)
• Got the core through Europractice
– FASTER (LPC Caen)
• ALTERA version OK
• Xilinx porting: synthesis OK; needs to be tested on real hardware
– PCI Express memory mapped (limited to ~30 MB/s)
• Phased delivery of hardware and firmware
Global Clock and trigger
• GTS Mezzanine Version 1
– OK, being used for testing the ATCA CARRIER
• GTS Mezzanine Version 2
– Prototype ready since one year
– Tests on hold awaiting end of ATCA carrier development
• Global trigger processor
– First version based on GTS mezzanines on a VME carrier (ready)
– Firmware developed in SystemC
– Automatic (almost) translation to VHDL
– Synthesized and running on a test board
• Timely completion of work problematic
Pulse Shape Analysis
• PSA algorithms validation
– Cologne experiment (2005)
– Liverpool scan-table data
– Imaging
– New data
– New experiment (only if major delay)
• Dependency on calculated signal basis is critical
• Narval implementation split in 3 blocks
– Readout and local event builder
– Preprocessing of data (calibrations,..)
• This is probably the most complicated part because we have to make in real time what took ~1 year for the Cologne experiment
– PSA: first implemented algorithm likely to be a Grid Search
• Need to finish the development of the AGATA Data Format (ADF) to inject proper data into the PSA
• Discussion on the needed hardware
c) Data Acquisition
The presentation by C Theisen was distributed by email prior to the meeting and is added to these minutes.
* Hardware
We found an agreement on the CPUs i.e. only new pizza boxes, Xeon.
Two option for the disks: Fiber Channel or DAS. Better performances with FC but more expensive.
Agreed to proceed to spend these funds with 5-6 pizza boxes and FC disks
Action: CT to organize orders with Ganil, Lyon and V Pucknell.
* Teams status
Orsay will visit Legnaro end of October to glue the Run Control + Narval + GUI
Hope to have a demo during the AGATA week.
d) Ancillary detectors and ancillary detector integration
Progress in the AGAVA firmware, problems found with the EXOGAM readout (VXI) in addition to the known GTS-AGAVA communication errors. New test programmed for the AGATA week ~12th to 16th November.
n.3 AGAVA modules for 9 k€ ordered by INFN-LNL to be delivered next year after prototype acceptance.
Design of the new PRISMA front-end readout hardware/software ongoing
Multiplicity Filter (Helena) as possible ancillary for the experimentation phase at LNL under evaluation.
e) Design and Infrastructure
From report of PR an email from John Strachan:
Mechanics meeting at IKP, Cologne
A meeting related to the mechanics of the AGATA frame and the assembly of the AGATA triple detectors in the Legnaro frame and the final spectrometer mechanical frame was initiated by John Strachan. Participants: John Strachan, Rob Griffiths STFC, Tobias Engert GSI, Bart Bruyneel, Juergen Eberth, Herbert Hess, Peter Reiter IKP, Cologne University, Heinz-Georg Thomas CTT
John Strachan distributed the following mail concerning the outcome of the meeting:
1. survey report.
1.1 The report showed that the faces of the endcap distorted significantly under
vacuum. This distortion is into the detector, and so is not cause the endcaps to collide,
but may be an issue for other reasons. Heinz-Georg has independently carried out some checks using DTIs (Dial Test Indicators), and also so some distortion. Heinz-Georg suggested glueing some stiffening rings around the inside of the endcap to reduce distortion if required.
At the cologne meeting we discussed:
1.2 The report showed that the mechanical construction of the detector was not good enough to achieve a sphere as required. Therefore an adjustment system will have to be designed and made. This was agreed by all.
2. A proposed detector adjustment frame was shown at the meeting, which involves a fixture
to locate the detector in its correct position and some adjustment for mounting. This design was accepted in principle and a prototype should be detailed and manufactured. (I will send an edrawing file on Monday. Some concern was raised as to the weight of the detector (80kg with adjustment rings), and so manual handling of the detector needs to considered.
3. Table top support frame - for use when maintaining the detector. The design shown at the
meeting was accepted with minor changes and should be detailed and manufactured.
4. HV boxes - no problems with space constraints.
There are some photos at the link also, I hope this helps.
Regards
John
Adjustment mechanics has become expensive.
What is the cost?
Action: JS will ask John Strachan for status.
Done. Cost to be discussed and action plan at next AMB.
Cost for adjustment is £4k per detector.
Need 1 gauge £5k
Setup for flange adjustment £4k
So for 6 detectors of the demonstrator £33k.
Also needed
Support stand £5k.
Lifting arrangement approx since not yet designed £10k
Grand total for demonstrator ~£50k (~75k euro).
The UK will cover these costs.
Not including the cost of ISEG modules the cost of the HV units will be 3k euro per crystal.
Action JS to organise sending the flange to LNL.
- LV modules have been ordered. Prototype to be delivered mid-Dec 07
- HV box: problems encountered at Saclay with this modules. Main problem is reliability and no current monitor nor current limitation possible. This is to be checked by GSI.
The Team has a VC tomorrow and will discuss the different ways to follow.
The document ISEG HV Module: problems and solutions, was distributed by G Duchene prior to the meeting.
- PLC and GUI: communications problems between GSI and Saclay. In addition, funding problems at GSI has lead to the GSI representation by J. Gerl at the last sitting meeting of the Team in Finland.
Saclay pays the travel for T. Habermann for a ~2 days stay at Saclay for a PLC-PC communication test with the right material and for a deep exchange about the GUI. Date not yet fixed but should be late Oct. or early Nov.
The document ISEG HV Module: problems and solutions, was distributed by G Duchene prior to the meeting.
Options were discussed. Tests are ongoing in GSI, should be done 3rd week of November.
Possibility of starting with standard SHV plugs and standard HV.
Low voltage module prototype ready mid Dec 2007. Want to do test of this with triple cluster and with one unit of the digitizer mid Jan 2008. If successful start production in Feb 2008.
Three first HV prototypes in March 2008.
Need to perform a test of HV boxes on a triple cluster in March 2008.
If a backup solution is required need to know where the equipment (HV) is. CAEN crate available in LNL but only negative polarity of the available HV cards are available. Does Köln have positive polarity HV cards ? Possibility to get HV modules from the spare parts of EUROBALL ?
Cabling of demonstrator: not all details clear right now; solutions to be found during installation. LNL working on this topic. LNL-GANIL meeting to be organised when a layout will be fixed.
Items that follow demonstrator to GANIL to be decided later on. Agreement between the three host- labs to be discussed and agreed.
f) Gamma-ray tracking, simulations and data analysis
GRID
– AGATA Virtual Organization
• Waiting for certificate, which is needed for submission of VO application. Certificate obtained 25/10!
• Approval of VO application should be fast
– Already existing AGATA GRID resources:
• Orsay:Tier2,4 CPUs, 3 TB
• Lyon:Tier3,20 CPUs, 10 TB + 5 TB
• Uppsala:Tierx (distr), 2 CPUs, 3 TB
– Need to start testing the GRID as soon as VO approved
– LNL (G. Maron) may host AGATA grid resources up to ~200 TB + necessary CPUs
AD data analysis funding: 20 k€
– French funds which must be spent latest end of Nov 2007 otherwise lost.
– Proposal of spending by O. Stezowski, Ch. Theisen et al. (discussed at recent French AGATA meeting and briefly with JN):
• Buy the minimum needed for LNL:
• 1 pizza box with ~2 TB of disks + at least 4 GB of RAM and few cores
• 1 or 2 cheap PCs for display
• Use remaining funds for the grid
Note: we do not have enough funding for offline storage and analysis
Discussed that this is not urgent and should not be spent yet.
Therefore, agreed that Lyon should purchase more DAQ storage. Expenditure on online analysis to be later when, for example GRID requirements are known.
Detector Database
The architecture of the detector database was presented.
Need to decide which database to use: CMS tracker database is proposed.
GD requests that each AMB member checks the architecture part in relation with its W Gr and comment as soon as possible.
AGATA week presentations
– Simulations:
• Enrico Farnea: Status report on simulations
• Grzegorz Jaworski: Simulations of ancillary devices: RFD, EXOGAM, DIAMANT
• Olivier Stezowski: Simulations of EXOGAM and AGATA